devbox
繁依Fanyi0
这个屌丝很懒，什么也没留下！
关注作者
热门标签
热门文章
当前位置:   article > 正文

ElasticSearch中全文搜索（单词搜索、多次搜索、组合搜索和权重搜索）

作者：繁依Fanyi0 | 2024-03-25 00:20:47

赞

踩

全文搜索

全文搜索两个最重要的方面是：

  • 相关性（Relevance） 它是评价查询与其结果间的相关程度，并根据这种相关程度对结果排名的一种能力，这种计算方式可以是 TF/IDF 方法、地理位置邻近、模糊相似，或其他的某些算法。
  • 分词（Analysis） 它是将文本块转换为有区别的、规范化的 token 的一个过程，目的是为了创建倒排索引以及 查询倒排索引。

一、构造数据

1、数据库中当前数据

 2、构建索引

  1. PUT http://127.0.0.1:9200/study
  2.  
  3. # 请求数据
  4. {
  5.     "settings": {
  6.         "index": {
  7.             "number_of_shards": "1",
  8.             "number_of_replicas": "0"
  9.         }
  10.     },
  11.     "mappings": {
  12.         "properties": {
  13.             "name": {
  14.                 "type": "text"
  15.             },
  16.             "age": {
  17.                 "type": "integer"
  18.             },
  19.             "mail": {
  20.                 "type": "keyword"
  21.             },
  22.             "hobby": {
  23.                 "type": "text",
  24.                 "analyzer": "ik_max_word"
  25.             }
  26.         }
  27.     }
  28. }
  29.  
  30. # 响应数据
  31. {
  32.     "acknowledged": true,
  33.     "shards_acknowledged": true,
  34.     "index": "study"
  35. }

3、添加数据

  1. POST http://127.0.0.1:9200/study/_bulk
  2.  
  3. # 请求数据
  4. {"index":{"_index":"study"}} 
  5. {"name":"张三","age": 20,"mail": "111@qq.com","hobby":"羽毛球、乒乓球、足球"} 
  6. {"index":{"_index":"study"}} 
  7. {"name":"李四","age": 21,"mail": "222@qq.com","hobby":"羽毛球、乒乓球、足球、篮球"} 
  8. {"index":{"_index":"study"}} 
  9. {"name":"王五","age": 22,"mail": "333@qq.com","hobby":"羽毛球、篮球、游泳、听音乐"} 
  10. {"index":{"_index":"study"}} 
  11. {"name":"赵六","age": 23,"mail": "444@qq.com","hobby":"跑步、游泳"} 
  12. {"index":{"_index":"study"}} 
  13. {"name":"孙七","age": 24,"mail": "555@qq.com","hobby":"听音乐、看电影"}
  14.  
  15. # 响应数据
  16. {
  17.     "took": 16,
  18.     "errors": false,
  19.     "items": [
  20.         {
  21.             "index": {
  22.                 "_index": "study",
  23.                 "_type": "_doc",
  24.                 "_id": "i6jJdoIBU4c5cKp3GGKx",
  25.                 "_version": 1,
  26.                 "result": "created",
  27.                 "_shards": {
  28.                     "total": 1,
  29.                     "successful": 1,
  30.                     "failed": 0
  31.                 },
  32.                 "_seq_no": 0,
  33.                 "_primary_term": 1,
  34.                 "status": 201
  35.             }
  36.         },
  37.         {
  38.             "index": {
  39.                 "_index": "study",
  40.                 "_type": "_doc",
  41.                 "_id": "jKjJdoIBU4c5cKp3GGKx",
  42.                 "_version": 1,
  43.                 "result": "created",
  44.                 "_shards": {
  45.                     "total": 1,
  46.                     "successful": 1,
  47.                     "failed": 0
  48.                 },
  49.                 "_seq_no": 1,
  50.                 "_primary_term": 1,
  51.                 "status": 201
  52.             }
  53.         },
  54.         {
  55.             "index": {
  56.                 "_index": "study",
  57.                 "_type": "_doc",
  58.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  59.                 "_version": 1,
  60.                 "result": "created",
  61.                 "_shards": {
  62.                     "total": 1,
  63.                     "successful": 1,
  64.                     "failed": 0
  65.                 },
  66.                 "_seq_no": 2,
  67.                 "_primary_term": 1,
  68.                 "status": 201
  69.             }
  70.         },
  71.         {
  72.             "index": {
  73.                 "_index": "study",
  74.                 "_type": "_doc",
  75.                 "_id": "jqjJdoIBU4c5cKp3GGKx",
  76.                 "_version": 1,
  77.                 "result": "created",
  78.                 "_shards": {
  79.                     "total": 1,
  80.                     "successful": 1,
  81.                     "failed": 0
  82.                 },
  83.                 "_seq_no": 3,
  84.                 "_primary_term": 1,
  85.                 "status": 201
  86.             }
  87.         },
  88.         {
  89.             "index": {
  90.                 "_index": "study",
  91.                 "_type": "_doc",
  92.                 "_id": "j6jJdoIBU4c5cKp3GGKx",
  93.                 "_version": 1,
  94.                 "result": "created",
  95.                 "_shards": {
  96.                     "total": 1,
  97.                     "successful": 1,
  98.                     "failed": 0
  99.                 },
  100.                 "_seq_no": 4,
  101.                 "_primary_term": 1,
  102.                 "status": 201
  103.             }
  104.         }
  105.     ]
  106. }

二、全文搜索

2.1、单词搜索

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "match": {
  7.             "hobby": "音乐"
  8.         }
  9.     },
  10.     "highlight": {
  11.         "fields": {
  12.             "hobby": {}
  13.         }
  14.     }
  15. }
  16.  
  17. # 响应数据
  18. {
  19.     "took": 67,
  20.     "timed_out": false,
  21.     "_shards": {
  22.         "total": 1,
  23.         "successful": 1,
  24.         "skipped": 0,
  25.         "failed": 0
  26.     },
  27.     "hits": {
  28.         "total": {
  29.             "value": 2,
  30.             "relation": "eq"
  31.         },
  32.         "max_score": 0.9395274,
  33.         "hits": [
  34.             {
  35.                 "_index": "study",
  36.                 "_type": "_doc",
  37.                 "_id": "j6jJdoIBU4c5cKp3GGKx",
  38.                 "_score": 0.9395274,
  39.                 "_source": {
  40.                     "name": "孙七",
  41.                     "age": 24,
  42.                     "mail": "555@qq.com",
  43.                     "hobby": "听音乐、看电影"
  44.                 },
  45.                 "highlight": {
  46.                     "hobby": [
  47.                         "听<em>音乐</em>、看电影"
  48.                     ]
  49.                 }
  50.             },
  51.             {
  52.                 "_index": "study",
  53.                 "_type": "_doc",
  54.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  55.                 "_score": 0.77041256,
  56.                 "_source": {
  57.                     "name": "王五",
  58.                     "age": 22,
  59.                     "mail": "333@qq.com",
  60.                     "hobby": "羽毛球、篮球、游泳、听音乐"
  61.                 },
  62.                 "highlight": {
  63.                     "hobby": [
  64.                         "羽毛球、篮球、游泳、听<em>音乐</em>"
  65.                     ]
  66.                 }
  67.             }
  68.         ]
  69.     }
  70. }

过程说明：

1. 检查字段类型

        爱好 hobby 字段是一个 text 类型（ 指定了IK分词器），这意味着查询字符串本身也应该被分词。

2. 分析查询字符串 。

        将查询的字符串 “音乐” 传入IK分词器中，输出的结果是单个项 音乐。因为只有一个单词项，所以 match 查询执 行的是单个底层 term 查询。

3. 查找匹配文档 。

        用 term 查询在倒排索引中查找 “音乐” 然后获取一组包含该项的文档，本例的结果是文档：3 、5 。

4. 为每个文档评分 。

        用 term 查询计算每个文档相关度评分 _score ，这是种将 词频（term frequency，即词 “音乐” 在相关文档的 hobby 字段中出现的频率）和 反向文档频率（inverse document frequency，即词 “音乐” 在所有文档的 hobby 字段中出现的频率），以及字段的长度（即字段越短相关度越高）相结合的计算方式。

2.2、单词搜索

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "match": {
  7.             "hobby": "音乐 篮球"
  8.         }
  9.     },
  10.     "highlight": {
  11.         "fields": {
  12.             "hobby": {}
  13.         }
  14.     }
  15. }
  16.  
  17. # 响应数据
  18. {
  19.     "took": 22,
  20.     "timed_out": false,
  21.     "_shards": {
  22.         "total": 1,
  23.         "successful": 1,
  24.         "skipped": 0,
  25.         "failed": 0
  26.     },
  27.     "hits": {
  28.         "total": {
  29.             "value": 3,
  30.             "relation": "eq"
  31.         },
  32.         "max_score": 1.5408251,
  33.         "hits": [
  34.             {
  35.                 "_index": "study",
  36.                 "_type": "_doc",
  37.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  38.                 "_score": 1.5408251,
  39.                 "_source": {
  40.                     "name": "王五",
  41.                     "age": 22,
  42.                     "mail": "333@qq.com",
  43.                     "hobby": "羽毛球、篮球、游泳、听音乐"
  44.                 },
  45.                 "highlight": {
  46.                     "hobby": [
  47.                         "羽毛球、<em>篮球</em>、游泳、听<em>音乐</em>"
  48.                     ]
  49.                 }
  50.             },
  51.             {
  52.                 "_index": "study",
  53.                 "_type": "_doc",
  54.                 "_id": "j6jJdoIBU4c5cKp3GGKx",
  55.                 "_score": 0.9395274,
  56.                 "_source": {
  57.                     "name": "孙七",
  58.                     "age": 24,
  59.                     "mail": "555@qq.com",
  60.                     "hobby": "听音乐、看电影"
  61.                 },
  62.                 "highlight": {
  63.                     "hobby": [
  64.                         "听<em>音乐</em>、看电影"
  65.                     ]
  66.                 }
  67.             },
  68.             {
  69.                 "_index": "study",
  70.                 "_type": "_doc",
  71.                 "_id": "jKjJdoIBU4c5cKp3GGKx",
  72.                 "_score": 0.77041256,
  73.                 "_source": {
  74.                     "name": "李四",
  75.                     "age": 21,
  76.                     "mail": "222@qq.com",
  77.                     "hobby": "羽毛球、乒乓球、足球、篮球"
  78.                 },
  79.                 "highlight": {
  80.                     "hobby": [
  81.                         "羽毛球、乒乓球、足球、<em>篮球</em>"
  82.                     ]
  83.                 }
  84.             }
  85.         ]
  86.     }
  87. }

        上面查询中只要是包含篮球和音乐的都被查询出来了。但是这有时候不能达到我们的要求，我们大部分时候都是希望两个词是同时包含的。这时候可以使用elasticsearch中指定词之间逻辑关系operator："and"

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "match": {
  7.             "hobby": {
  8.                 "query": "音乐 篮球",
  9.                 "operator": "and"
  10.             }
  11.         }
  12.     },
  13.     "highlight": {
  14.         "fields": {
  15.             "hobby": {}
  16.         }
  17.     }
  18. }
  19.  
  20. # 响应结果
  21. {
  22.     "took": 6,
  23.     "timed_out": false,
  24.     "_shards": {
  25.         "total": 1,
  26.         "successful": 1,
  27.         "skipped": 0,
  28.         "failed": 0
  29.     },
  30.     "hits": {
  31.         "total": {
  32.             "value": 1,
  33.             "relation": "eq"
  34.         },
  35.         "max_score": 1.5408251,
  36.         "hits": [
  37.             {
  38.                 "_index": "study",
  39.                 "_type": "_doc",
  40.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  41.                 "_score": 1.5408251,
  42.                 "_source": {
  43.                     "name": "王五",
  44.                     "age": 22,
  45.                     "mail": "333@qq.com",
  46.                     "hobby": "羽毛球、篮球、游泳、听音乐"
  47.                 },
  48.                 "highlight": {
  49.                     "hobby": [
  50.                         "羽毛球、<em>篮球</em>、游泳、听<em>音乐</em>"
  51.                     ]
  52.                 }
  53.             }
  54.         ]
  55.     }
  56. }

        上面的测试结果都是选择了"and"和"or"两个极端情况下，但是在真正搜索中，我们不会使用这两个极端情况的，这样就需要另外一种查询方式，即为只需要符合一定的相似度就可以查询到的数据，在elasticsearch中就支持这种查询方式，如使用minimum_should_match来指定匹配度，如60%。

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "match": {
  7.             "hobby": {
  8.                 "query": "游泳 羽毛球",
  9.                 "minimum_should_match": "80%"
  10.             }
  11.         }
  12.     },
  13.     "highlight": {
  14.         "fields": {
  15.             "hobby": {}
  16.         }
  17.     }
  18. }
  19.  
  20. # 响应数据
  21. {
  22.     "took": 6,
  23.     "timed_out": false,
  24.     "_shards": {
  25.         "total": 1,
  26.         "successful": 1,
  27.         "skipped": 0,
  28.         "failed": 0
  29.     },
  30.     "hits": {
  31.         "total": {
  32.             "value": 3,
  33.             "relation": "eq"
  34.         },
  35.         "max_score": 2.1933634,
  36.         "hits": [
  37.             {
  38.                 "_index": "study",
  39.                 "_type": "_doc",
  40.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  41.                 "_score": 2.1933634,
  42.                 "_source": {
  43.                     "name": "王五",
  44.                     "age": 22,
  45.                     "mail": "333@qq.com",
  46.                     "hobby": "羽毛球、篮球、游泳、听音乐"
  47.                 },
  48.                 "highlight": {
  49.                     "hobby": [
  50.                         "<em>羽毛球</em>、篮球、<em>游泳</em>、听音乐"
  51.                     ]
  52.                 }
  53.             },
  54.             {
  55.                 "_index": "study",
  56.                 "_type": "_doc",
  57.                 "_id": "i6jJdoIBU4c5cKp3GGKx",
  58.                 "_score": 1.7171206,
  59.                 "_source": {
  60.                     "name": "张三",
  61.                     "age": 20,
  62.                     "mail": "111@qq.com",
  63.                     "hobby": "羽毛球、乒乓球、足球"
  64.                 },
  65.                 "highlight": {
  66.                     "hobby": [
  67.                         "<em>羽毛球</em>、乒乓<em>球</em>、足球"
  68.                     ]
  69.                 }
  70.             },
  71.             {
  72.                 "_index": "study",
  73.                 "_type": "_doc",
  74.                 "_id": "jKjJdoIBU4c5cKp3GGKx",
  75.                 "_score": 1.6262295,
  76.                 "_source": {
  77.                     "name": "李四",
  78.                     "age": 21,
  79.                     "mail": "222@qq.com",
  80.                     "hobby": "羽毛球、乒乓球、足球、篮球"
  81.                 },
  82.                 "highlight": {
  83.                     "hobby": [
  84.                         "<em>羽毛球</em>、乒乓<em>球</em>、足球、篮球"
  85.                     ]
  86.                 }
  87.             }
  88.         ]
  89.     }
  90. }

2.3、组合搜索

        在搜索时除了上面的方法外，还可以使用过滤器中的bool组合搜索。

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "bool": {
  7.             "must": {
  8.                 "match": {
  9.                     "hobby": "篮球"
  10.                 }
  11.             },
  12.             "must_not": {
  13.                 "match": {
  14.                     "hobby": "音乐"
  15.                 }
  16.             },
  17.             "should": [
  18.                 {
  19.                     "match": {
  20.                         "hobby": "游泳"
  21.                     }
  22.                 }
  23.             ]
  24.         }
  25.     },
  26.     "highlight": {
  27.         "fields": {
  28.             "hobby": {}
  29.         }
  30.     }
  31. }
  32.  
  33. # 响应数据
  34. {
  35.     "took": 6,
  36.     "timed_out": false,
  37.     "_shards": {
  38.         "total": 1,
  39.         "successful": 1,
  40.         "skipped": 0,
  41.         "failed": 0
  42.     },
  43.     "hits": {
  44.         "total": {
  45.             "value": 1,
  46.             "relation": "eq"
  47.         },
  48.         "max_score": 0.77041256,
  49.         "hits": [
  50.             {
  51.                 "_index": "study",
  52.                 "_type": "_doc",
  53.                 "_id": "jKjJdoIBU4c5cKp3GGKx",
  54.                 "_score": 0.77041256,
  55.                 "_source": {
  56.                     "name": "李四",
  57.                     "age": 21,
  58.                     "mail": "222@qq.com",
  59.                     "hobby": "羽毛球、乒乓球、足球、篮球"
  60.                 },
  61.                 "highlight": {
  62.                     "hobby": [
  63.                         "羽毛球、乒乓球、足球、<em>篮球</em>"
  64.                     ]
  65.                 }
  66.             }
  67.         ]
  68.     }
  69. }

注意：上面示例中在搜索结果中必须包含篮球，不能包含音乐，如果包含了游泳，那么它的相似度更高。

评分的计算规则

  • bool 查询会为每个文档计算相关度评分 _score ， 再将所有匹配的 must 和 should 语句的分数 _score 求和， 最后除以 must 和 should 语句的总数。
  • must_not 语句不会影响评分； 它的作用只是将不相关的文档排除。

注意：默认情况下，should中的内容不是必须匹配的，如果查询语句中没有must，那么就会至少匹配其中一个。当然了， 也可以通过minimum_should_match参数进行控制，该值可以是数字也可以的百分比。

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "bool": {
  7.             "should": [
  8.                 {
  9.                     "match": {
  10.                         "hobby": "游泳"
  11.                     }
  12.                 },
  13.                 {
  14.                     "match": {
  15.                         "hobby": "篮球"
  16.                     }
  17.                 },
  18.                 {
  19.                     "match": {
  20.                         "hobby": "音乐"
  21.                     }
  22.                 }
  23.             ],
  24.             "minimum_should_match": 2
  25.         }
  26.     },
  27.     "highlight": {
  28.         "fields": {
  29.             "hobby": {}
  30.         }
  31.     }
  32. }
  33.  
  34. # 响应数据
  35. {
  36.     "took": 4,
  37.     "timed_out": false,
  38.     "_shards": {
  39.         "total": 1,
  40.         "successful": 1,
  41.         "skipped": 0,
  42.         "failed": 0
  43.     },
  44.     "hits": {
  45.         "total": {
  46.             "value": 1,
  47.             "relation": "eq"
  48.         },
  49.         "max_score": 2.3112378,
  50.         "hits": [
  51.             {
  52.                 "_index": "study",
  53.                 "_type": "_doc",
  54.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  55.                 "_score": 2.3112378,
  56.                 "_source": {
  57.                     "name": "王五",
  58.                     "age": 22,
  59.                     "mail": "333@qq.com",
  60.                     "hobby": "羽毛球、篮球、游泳、听音乐"
  61.                 },
  62.                 "highlight": {
  63.                     "hobby": [
  64.                         "羽毛球、<em>篮球</em>、<em>游泳</em>、听<em>音乐</em>"
  65.                     ]
  66.                 }
  67.             }
  68.         ]
  69.     }
  70. }

2.4、权重搜索

        在一些情况下，可能会对某些词增加权重来影响该条数据的得分。

  1. POST  http://127.0.0.1:9200/study/_search
  2.  
  3. # 请求数据
  4. {
  5.     "query": {
  6.         "bool": {
  7.             "must": {
  8.                 "match": {
  9.                     "hobby": {
  10.                         "query": "游泳篮球",
  11.                         "operator": "and"
  12.                     }
  13.                 }
  14.             },
  15.             "should": [
  16.                 {
  17.                     "match": {
  18.                         "hobby": {
  19.                             "query": "音乐",
  20.                             "boost": 10
  21.                         }
  22.                     }
  23.                 },
  24.                 {
  25.                     "match": {
  26.                         "hobby": {
  27.                             "query": "跑步",
  28.                             "boost": 2
  29.                         }
  30.                     }
  31.                 }
  32.             ]
  33.         }
  34.     },
  35.     "highlight": {
  36.         "fields": {
  37.             "hobby": {}
  38.         }
  39.     }
  40. }
  41.  
  42. # 响应数据
  43. {
  44.     "took": 5,
  45.     "timed_out": false,
  46.     "_shards": {
  47.         "total": 1,
  48.         "successful": 1,
  49.         "skipped": 0,
  50.         "failed": 0
  51.     },
  52.     "hits": {
  53.         "total": {
  54.             "value": 1,
  55.             "relation": "eq"
  56.         },
  57.         "max_score": 9.24495,
  58.         "hits": [
  59.             {
  60.                 "_index": "study",
  61.                 "_type": "_doc",
  62.                 "_id": "jajJdoIBU4c5cKp3GGKx",
  63.                 "_score": 9.24495,
  64.                 "_source": {
  65.                     "name": "王五",
  66.                     "age": 22,
  67.                     "mail": "333@qq.com",
  68.                     "hobby": "羽毛球、篮球、游泳、听音乐"
  69.                 },
  70.                 "highlight": {
  71.                     "hobby": [
  72.                         "羽毛球、<em>篮球</em>、<em>游泳</em>、听<em>音乐</em>"
  73.                     ]
  74.                 }
  75.             }
  76.         ]
  77.     }
  78. }

声明：本文内容由网友自发贡献，不代表【wpsshop博客】立场，版权归原作者所有，本站不承担相应法律责任。如您发现有侵权的内容，请联系我们。转载请注明出处：https://www.wpsshop.cn/w/繁依Fanyi0/article/detail/305755
推荐阅读
相关标签

Copyright © 2003-2013 www.wpsshop.cn 版权所有，并保留所有权利。

  

闽ICP备14008679号